Track: Modeling and Simulation

Abstract

This paper describes the effects of chromosome length for solving maritime inventory routing problems (MIRP) by using a hyper-heuristics based Genetic Algorithm (GA). The approach uses a set of heuristic combinations, each of which consist of strategies that correspond to a ship assignment. These strategies are represented by a chromosome that may have several assignments. We examine several number of chromosome length to encourage the evolution of good heuristics combinations. Moreover, a variation of several number chromosome length is necessary since we do not know in advance how many ship assignments are needed to cover demands during a predefined planning horizon. At every iteration a number of chromosomes are evaluated and evolved within a GA framework. In this study, the approach has been applied on several test cases for transporting multiple oil products from a production facility to some consumption ports, by using several heterogeneous ships with undedicated compartments. The results show that a hyper-heuristics based GA reaches the same global optimal as the solutions in the mathematical model, but with a significant decrease in computational time. Moreover, the use two numbers of chromosome length proves that three assignments in one step (3AiOS) mostly got better solutions and lower minimum total number of assignments than the two assignments (2AiOS).